128
Remanufacturing and Advanced Machining
the Internet of Manufacturing Services (IoMs), the Internet of People (IoP), an
embedded system, and Integration of Information and Communication Technology
(IICT). The three key features of the IoT are context, omnipresence, and optimization (Vaidiya et al., 2018).
Phuyal et al. (2020) emphasize that the use of the IoT has been adopted widely in
the field of industrial automation in developed countries. The architecture of IoT is
basically divided into four layers:
1. The first layer includes sensors and actuators which are integrated into
hardware to collect information.
2. The networking layer serves transmission of collected data from sensors to
control units and from control units to actuators.
3. To access services based on user needs and to interact with the control
units, a service and interface layer is distinguished and defined. Information
from a physical system is collected through the use of sensors and different
machine learning tools. The collected information is then managed systematically, processed in local smart devices, and then sent to the application
layer through the network layer for implementation in the respective fields.
4. The application layer decides to take necessary action in an IoT integrated
system (Phuyal et al., 2020).
The term “cyber–physical system” (CPS) can be understood as a system where
natural and human-made systems (physical space) are tightly integrated into computation, communication, and control systems (cyber space). Decentralization and
autonomous behavior of the production process are the main features of the CPS,
dependent on adoption and reconfiguration of product structure supply networks. The
latter may be considered as collaborative cyber-physical systems used in manufacturing systems. Continuous intelligent interchanging of data is carried out by linking
cyber-physical systems through cloud systems in real time. The basic requirement of
real-time-oriented manufacturing operation and optimization of actual production
systems is achieved by means of proper sensors in the CPS (Vaidiya et al., 2018).
A digital twin (DT) can be created as a result of the fusion of real and digital systems. When field data are used or transferred to a simulation model, a realistic digital
image is created; validation and retrofitting of the predictive model are performed
with realistic data. The simulation model learns from actual experiences and events
in reality, which makes it no longer a simple theoretical calculation. This model
reproduces the real behavior of a machine or individual components almost identically and is thus called “a digital twin” (Werner et al., 2019).
Next-generation artificial intelligence (Next-Gen AI) is a framework that incorporates explainability, interpretability, transfer learning, and ensemble learning with
a wide variety of learning architectures. It uses the context of previously known
information and facilitates the use of human knowledge and experience for experimental trial design and interpretation of results. Since the first-generation AI is
being used in surveying and classifying omics data, it is designed to solve welldefined tasks of single-omics datasets and does not require the integration of data
Remanufacturing and Advanced Machining
the Internet of Manufacturing Services (IoMs), the Internet of People (IoP), an
embedded system, and Integration of Information and Communication Technology
(IICT). The three key features of the IoT are context, omnipresence, and optimization (Vaidiya et al., 2018).
Phuyal et al. (2020) emphasize that the use of the IoT has been adopted widely in
the field of industrial automation in developed countries. The architecture of IoT is
basically divided into four layers:
1. The first layer includes sensors and actuators which are integrated into
hardware to collect information.
2. The networking layer serves transmission of collected data from sensors to
control units and from control units to actuators.
3. To access services based on user needs and to interact with the control
units, a service and interface layer is distinguished and defined. Information
from a physical system is collected through the use of sensors and different
machine learning tools. The collected information is then managed systematically, processed in local smart devices, and then sent to the application
layer through the network layer for implementation in the respective fields.
4. The application layer decides to take necessary action in an IoT integrated
system (Phuyal et al., 2020).
The term “cyber–physical system” (CPS) can be understood as a system where
natural and human-made systems (physical space) are tightly integrated into computation, communication, and control systems (cyber space). Decentralization and
autonomous behavior of the production process are the main features of the CPS,
dependent on adoption and reconfiguration of product structure supply networks. The
latter may be considered as collaborative cyber-physical systems used in manufacturing systems. Continuous intelligent interchanging of data is carried out by linking
cyber-physical systems through cloud systems in real time. The basic requirement of
real-time-oriented manufacturing operation and optimization of actual production
systems is achieved by means of proper sensors in the CPS (Vaidiya et al., 2018).
A digital twin (DT) can be created as a result of the fusion of real and digital systems. When field data are used or transferred to a simulation model, a realistic digital
image is created; validation and retrofitting of the predictive model are performed
with realistic data. The simulation model learns from actual experiences and events
in reality, which makes it no longer a simple theoretical calculation. This model
reproduces the real behavior of a machine or individual components almost identically and is thus called “a digital twin” (Werner et al., 2019).
Next-generation artificial intelligence (Next-Gen AI) is a framework that incorporates explainability, interpretability, transfer learning, and ensemble learning with
a wide variety of learning architectures. It uses the context of previously known
information and facilitates the use of human knowledge and experience for experimental trial design and interpretation of results. Since the first-generation AI is
being used in surveying and classifying omics data, it is designed to solve welldefined tasks of single-omics datasets and does not require the integration of data